RE: NEC-LIST: Definition of Gain

----------
From: Matt Taylor[SMTP:matayl_at_INR.Net]
Sent: Friday, March 03, 2000 6:18 PM
To: nec-list_at_ee.ubc.ca
Cc: g.hagn_at_ieee.org
Subject: Re: NEC-LIST: Definition of Gain

George,

I see that everyone disagrees with me on the question of lumping
impedance mismatch loss and polarization loss in with gain. Let me
give an example to try to support my contention that the definition of
antenna gain should include both mismatch losses.
        :
        :
Matt Taylor
----------

I agree with Matt, at least for my situation. I can see how others
who use antennas differently would find the current definition more
useful. I primarily use an antenna as an RF delivery device
connecting to a fixed-impedance system.

1. In comparing antennas with different impedance, directivity,
efficiency, and polarization characteristics, I primarily care how
much energy of a given polarization the antenna will deliver to my
system. An antenna with high thermal efficiency and high directivity
is not much use to me if its impedance is wildly mismatched to my
system or if its specified response is to varying polarizations. I
typically work with broad-band systems where tuning is not an option
and 6:1 VSWR over parts of the band are not uncommon, so impedance
mismatch is can be more of a factor in range performance than
efficiency.

In the case where a tuner it to be employed, efficiency and
directivity are, of course, more useful in defining the interface
between the aperture and the tuner. Our systems folks often define an
"antenna" to include the tuner. Again, between the tuner and the
fixed-impedance system the power-delivery quantity which includes
mismatch losses (the quantity formerly known as "gain") is most
useful.

2. Antenna gain measurements are typically performed with fixed
impedance systems with fixed-polarization sources (at least in my
experience). If we wished to report the IEEE definition of gain based
on test results we would have to either tweak impedance and
polarization at every test point, or adjust measured values upward
after testing. Either operation puts more burden on the testers (and
frankly, requires more faith) than is reasonable in many situations,
especially since the directly-measured raw data is what I am most
interested in the first place. In terms of polarization, though, I
really did think that dBil referred to a fixed linear polarization,
defined by the test coordinate system. Is that not correct?

The debate seems to boil down to the intended use of the data. A
gain-like quantity which includes mismatch and polarization loss boils
everything down to one directly-measurable quantity for use in a
fixed-impedance system. Gain which includes only directivity and
efficiency is more useful for integration into variable-impedance
systems (such as tuning-stages). Rather than change long-standing
definitions, I guess that we will simply have be extremely clear on
our use of the terms involved. Instead of specifying a 3 dBil
antenna, I should say something like "3 dBi_VP_50_Ohm."

        David de Schweinitz
Received on Tue Mar 07 2000 - 12:56:11 EST